[OE-core] precise details on processing of multiple .bbappend files across multiple layers?
Christopher Larson
clarson at kergoth.com
Fri Apr 22 23:18:27 UTC 2016
On Fri, Apr 22, 2016 at 3:48 PM Robert P. J. Day <rpjday at crashcourse.ca>
wrote:
>
> i realize i should know the answer to this, but it's friday and i'm
> tired. i want to know the precise processing order for the bbappend
> files for a single recipe over multiple layers. (i realize it's
> already a bad design if i have so many bbappend files for the same
> recipe that i get confused knowing how they're processed. :-)
>
> if i have a recipe, say, rday_1.2.3.bb, and numerous rday*bbappend
> files spread across several layers, i already know that those layers
> are processed in increasing priority, so that higher-priority layers
> will overwrite the settings of lower-priority layers. so far, so good.
>
> now, in each layer, the selection of any bbappend file will be based
> on specificity, so in my case, selection order might be:
>
> * rday_1.2.3.bbappend
> * rday_1.%.bbappend
> * rday_%.bbappend
>
> and as long as that's true, i'm assuming that search order is
> restarted for each layer, so that if i have:
>
> * layer priority 5, rday_1.2.%.bbappend
> * layer priority 6, rday_1.%.bbappend
>
> the first of the above will be processed, but will be superseded by
> the second even though it is a more generic wildcard, correct?
>
Yes, the specificity of the append isn't a factor in ordering today, as far
as I know.
> finally, if i have multiple layers with the same priority, what is
> the processing order of bbappend files there? is it random? based on
> wildcard specificity?
>
You should treat it as undefined. Don't rely on it. I *think* a recent
patch went in to at least make it deterministic by sorting in some fashion,
but afaik that's not guaranteed, and wasn't true in the past.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openembedded.org/pipermail/openembedded-core/attachments/20160422/0e5d42ab/attachment-0002.html>
More information about the Openembedded-core
mailing list